Journal Number 110
November 2008

ORIGINAL PAPERS

Chromosomes of New Zealand Native Orchids - Part 2
By Murray Dawson, Landcare Research, Lincoln

Part 1 discussed newly published chromosome counts of the New Zealand native orchids, and
the discovery of polyploid, allopolyploid, and aneuploid species [1].

In this second part, I change focus to consider the higher level (genus and above) implications of
this work. These chromosome counts provide a valuable new set of characters that seem remarkably informative for the taxonomy (classification) and phylogeny (evolutionary reconstructions) of the
New Zealand and Australian orchids.

Both articles summarise a paper that I published with my co-authors Brian Molloy and Ernst
Beuzenberg in the December 2007 issue of the New Zealand Journal of Botany (NZJB) [2].

The Genus Problem

Readers of the NZNOG journal will be well-aware of the extraordinary number of name changes
affecting the New Zealand orchids, especially at the genus level.

Many names have been reinstated or segregated out of long-standing and well-known Australasian
orchid genera. For example, David Jones and Mark Clements of Canberra (and their co-authors)
[3, 4] split Corybas (the spider and helmet orchids) into segregate genera including Anzybas,
Nematoceras, and the New Zealand endemic genera Molloybas and Singularybas.

Another well-known example is Pterostylis (the greenhood orchids), split into the segregate genus
Plumatichilos by Dariusz Szlachetko of Poland [5], and then further subdivided by Jones and
Clements into 15 segregate genera including Diplodium, Hymenochilus and Linguella [6, 7, 8].

To summarise some other changes:

• Acianthus was split into Cyrtostylis and Townsonia;
• Bulbophyllum was split into several genera including Adelopetalum and Ichthyostomum;
• Caladenia was split into several genera including Petalochilus and Stegostyla;
• Caleana/Paracaleana minor was dropped in favour of the reinstated genus Sullivania;
• Chiloglottis was split into Myrmechila and Simpliglottis;
• Dendrobium cunninghamii was transferred to a new genus, Winika;
• Lyperanthus antarcticus was transferred to a new genus, Waireia (as W. stenopetala);
• Prasophyllum nudum and P. pumilum were transferred to Genoplesium and then Corunastylis;
• Yoania australis was transferred to a new genus, Danhatchia.

This multitude of proposed changes, compounded by a lack of consensus, has made it difficult to
know what names to use. There are frequent references to this problem within the NZNOG journal
[e.g., NZNOG 51:23; 92:7; 94:22; 97:24, 29; 102:46].

While some of the changes now seem to make sense and are generally accepted by most people
(e.g., usage of Danhatchia and Waireia), others are complex and remain hotly debated.

A prime example of extensive and contentious changes is the range of genus-level treatments of
Caladenia and its segregates. When commenting on the earlier work of Szlachetko [5], David Jones
and his colleagues stated: "His treatment, which appears to be based on a very limited range of
herbarium specimens, is superficial, skeletal and indicates a lack of basic knowledge of the orchid
groups involved" [9]. And they continue with other strong words. Hopper and Brown [10], in turn,
also revise Caladenia, but argue for a conservative treatment and criticise both earlier works [5, 9].
And so the debate continues.

For New Zealand, Ian St George provides annual lists of the names that he accepts ("a personal
opinion, wrested from observation, discussion, plagiarism and taxonomic punch-ups"!) [e.g., 11].

For our NZJB paper [2] we follow the taxonomic treatments of our co-author Brian Molloy. Brian
works closely with his Australian colleagues, David Jones and Mark Clements, and together they
have proposed the majority of the genus-level changes.

Characters and Classification Systems

Traditional taxonomic characters include vegetative (e.g., leaf size and shape) and reproductive
(e.g., flower structures) characters. Because most traditional characters are relatively easy to
observe and measure, they are known as macromorphological characters. These visible characters
form the basis of traditional classifications and remain important to this day.

Classification systems are extensive hierarchies that operate not only at the genus and species level,
but at higher (and lower) taxonomic levels too. For example, following one system, the ladies' tresses
orchid Spiranthes novae-zelandiae belongs to subtribe Spiranthinae, which in turn is a member of
tribe Cranichideae, of subfamily Orchidoideae, of family Orchidaceae.

Traditional classifications of the orchids have relied on a few key floral characters, such as anther
configuration and column structure. This early reliance on floral characters created some artificial
groupings due to parallel and convergent evolution well-known in the orchids. For the orchids, the
best-known classification systems are those of Robert Dressler of the Missouri Botanical Garden.
His most recent [12] was published in 1993, is still widely used, and includes discussion of other
characters, including pollen, seed, and anatomical features to supplement traditional floral characters.

The next step in attempting to create a more natural (phylogenetic) classification system has been
taken at the DNA level. From the early 1990s DNA sequencing has revolutionised our understanding
of the relationships of plant groups. These invisible but highly informative characters are used to
create what are called phylogenetic trees - reconstructions of the evolutionary relationships of species.

Orchid phylogenies were incorporated by Mark Chase of Kew and his colleagues into a new
classification [13], published 10 years after Dressler [12]. The authors acknowledge that their 2003
classification is not the "final word", but it has been largely followed in the Genera Orchidacearum
series. Production of this magnificent series is also based at Kew and coordinated by Alec Pridgeon;
the editors have currently completed Vol. 4 [14] in a six-volume set.

However, molecular studies are open to interpretation and have not resolved all of the problems.
Between studies there are differences in the gene regions used and the number of species sampled
which may also produce differing results. For example, in the same year as Mark Chase et al's.
classification [13], David Jones and his colleagues arrived at a markedly different arrangement for
the Tribe Diurideae [15], yet both classifications used DNA sequences.

For our 2007 paper [2], we decided that a mixed classification system would best suit our needs.
For the most part, we followed the system of Chase et al. [13], because it provided comprehensive
coverage and incorporated the recent molecular work. However, for the tribe Diurideae, we followed
the classification of Clements et al. [3] because it provided the best (but not a perfect) fit with the
chromosome information. Table 1 shows our mixed classification, along with a summary of
genus-level chromosome counts made by us and other workers. To show the full range of variation,
chromosome numbers from genera shared by New Zealand, Australia, and other countries are
included in the table (below)

Chromosome Evidence

Size, shape, and number of chromosomes provide valuable characters independent of
macromorphological and DNA data.

As mentioned previously [1, 2], chromosome counts of Australian representatives are still limited,
and chromosome evidence on its own, particularly at higher taxonomic levels, has to be treated
cautiously. However, what was surprising for the orchids was that for several groups the
chromosome evidence strongly supports the recognition of at least some segregates (e.g., for
some of the segregate genera of Caladenia, Corybas, Prasophyllum, and Pterostylis) and their
subtribal placements.

We also found examples where none of the taxonomic treatments fully fitted the chromosome
evidence. Some genera and subtribes remained chromosomally heterogeneous despite the fact
that they have recently been revised. These may not be natural groups and may require further
taxonomic revision and corroborative chromosome counts (e.g., within Prasophyllum as currently
circumscribed).

What follows are some of the interesting case studies among New Zealand and related genera.

Subtribe Pterostylidinae: There is good chromosome support for segregating Pterostylis
[7, 8]. Within the narrower definition of Pterostylis, subgenus Pterostylis is characterised by
2n = 42 (admittedly based on limited counts) whereas the other two subgenera share 2n = 44.

In contrast, the segregate genera Diplodium and Linguella are centred on 2n = 50, Hymenochilus
have a range of counts, 2n = c. 48, 52, 54, & 62, and Plumatichilos has 2n = 50-54.

Tribe Diurideae: The subtribes and genera that follow are all in tribe Diurideae (see Table 1
for an overview). As previously mentioned, we follow David Jones and Mark Clements treatment
of this tribe [3, 15, 16] because they provide a better fit with our chromosome data compared
with Chase et al.'s classification [13]. Jones and Clements split the tribe more finely and in many
cases their subtribes are supported by differences in chromosome number and/or chromosome
morphology.

Subtribe Acianthinae: Acianthus has 2n = 40 and 2n = c. 60 whereas Cyrtostylis has 2n = 44-46;
chromosomes of the other non-New Zealand genera of the Acianthus alliance are unknown.
Although the chromosome numbers of Acianthus and Cyrtostylis differ, the chromosome size and
morphology is similar. In contrast, chromosomes of the Acianthus alliance are smaller than those
of the Corybas alliance.

There is support for at least some of the segregate genera of Corybas where they have different
chromosome numbers. All share similar chromosome morphology indicating that they remain a
relatively closely related group. Molloybas and Singularybas share 2n = 34, presumably derived
through aneuploidy (loss of one chromosome pair) from 2n = 36 found elsewhere in the Corybas
alliance. Anzybas has 2n = 36, the same as Nematoceras except for the N. trilobum agg., which
has both 2n = 36 (diploid) and 2n = 72 (tetraploid) representatives. Interestingly, Corybas in the
restricted sense [4] differs with 2n = 54+2 (or 2n = 56) chromosomes.

Subtribe Adenochilidinae: This subtribe contains only one genus, Adenochilus, and was
recently separated from subtribe Caladeniinae [15]. We counted one of the two species,
Adenochilus gracilis, with 2n = 38, a number that differs from those known in Jones et al.'s
concept of Caladeniinae [15].

Subtribe Caladeniinae: This subtribe has had major recircumscriptions [3, 17], where 10
segregate genera were recognised from Caladenia, and several other genera were transferred
to the new subtribes Adenochilidinae and Megastylidinae. There is strong chromosome support
for some of the segregate genera, which may help settle the debate surrounding these
recircumscriptions.

We counted New Zealand species from two segregates of Caladenia, namely Petalochilus and
Stegostyla. Although Petalochilus has 2n = 40 (with some aneuploidy), there are two different
chromosome types (cytotypes) within this genus, suggesting that it may not be a natural group.

One cytotype has small chromosomes (P. aff. carneus and P. minor with 2n = 40) and the other
has chromosomes with different morphology and of a larger size (seen in only P. chlorostylus
with 2n = 39, 40, 41). Further work is needed to reconcile these differences between chromosome
number and the existing taxonomy.

Stegostyla lyallii of New Zealand has 2n = 47, 48, but again, further chromosome counts are needed
of other species to determine if this is a consistent difference that supports the separation of
Stegostyla at the genus level.

Peakall and James [18] counted chromosomes of several Australian orchids. Updating the names
that they used with the subsequent generic recircumscriptions of Caladenia suggests that there is
chromosome support for some of the other segregate genera in Australia. Caladenia (in the restricted
sense) has 2n = 48, Arachnorchis has mainly 2n = 44, Leptoceras has 2n = 44, and Jonesiopsis has
2n = 46. Our counts, in conjunction with those of Peakall and James [18], appear to form an aneuploid
series that characterises and strongly supports many of the segregate genera in subtribe Caladeniinae.

Subtribe Cryptostylidinae: Following Clements et al. [3], this subtribe is made up of only one
genus, Cryptostylis, but there is nevertheless an interesting range of chromosome numbers.

Cryptostylis subulata from New Zealand has 2n = 64 [2] whereas two Australian species have
2n = 56 and a high polyploid count of 2n = c. 187 [18]. Also, one species from Thailand has
2n = 42 [19]. Like so many other groups, further chromosome counts are needed.

Subtribe Diuridinae: Diuris and Orthoceras are the only genera that Jones and Clements recognise
in this subtribe [3, 15, 16], although Dressler [12] and Chase et al. [13] included Epiblema.

Chromosomally, Diuris and Orthoceras are quite different from one another even though they
constitute a well-supported group in a molecular analysis [3].

Orthoceras has about six species but several are undescribed. We counted two species; most have
2n = 42, but we also obtained 2n = 40 and 2n = 44. The chromosomes were small and it is uncertain
if this narrow range of numbers represents real variation. Ours are the only known chromosome
counts for this genus.

Diuris is absent from New Zealand and centred in Australia. In contrast to Orthoceras, Diuris is
more species-rich and has much larger chromosomes and a wider range of numbers - various
species have 2n = 34, 36, 36-38 [2], and 2n = 38, 56, c. 112 [18].

Subtribe Drakaeinae: Myrmechila and Simpliglottis are two segregate genera of Chiloglottis [5, 20]. Myrmechila trapeziformis (from both sides of the Tasman) and Simpliglottis (S. cornuta and S. valida counted from New Zealand) have 2n = 40 with similar chromosomes.

Sullivania minor differs markedly with 2n = 54(+2). Sullivania is a recently reinstated genus [20].

The only other chromosome counts in subtribe Drakaeinae are for two Australian genera and species (Spiculaea ciliata, 2n = c. 40; Drakaea glyptodon, 2n = c. 44) [18]. Chromosome counts are lacking in this subtribe and Chiloglottis in the restricted sense [20] does not appear to have been counted yet (Fig. 1).

Subtribe Megastylidinae: This new subtribe was created by Jones et al. [15] to accommodate eight genera formerly placed in subtribes Caladeniinae (Aporostylis) and Thelymitrinae (Burnettia, Leporella, Lyperanthus, Megastylis, Pyrorchis, Rimacola, and Waireia).

Chromosome evidence supports removal of at least some genera from their previous placements
in other subtribes, but clearly divides the Megastylidinae into at least two (probably three) groups
(Fig. 1). The first group has 2n = 40 moderately small chromosomes and is represented by
Aporostylis and Waireia of New Zealand. Waireia stenopetala used to be placed in Lyperanthus
(as L. antarcticus). This earlier placement in Lyperanthus was quite wrong as the chromosomes
are completely different - we knew this well before Waireia was created in 1997 [21].

The second well-defined group has the longest orchid chromosomes that I have examined.

This group is represented by the Australian Lyperanthus suaveolens (2n = 44) [2] and Pyrorchis
(2n = 42) [18]. The chromosome number of Pyrorchis may be derived through aneuploid reduction
of a 2n = 44 ancestor; the Aporostylis and Waireia cytotype, characterised by 2n = 40, probably has
an independent origin.

Like Lyperanthus, two other non-New Zealand genera have 2n = 44; Leporella fimbriata of Australia
[18] and Megastylis gigas of New Caledonia [22]. However, the published chromosome illustrations
appear to show only moderately sized chromosomes, so maybe there is a third group within the
Megastylidinae.

Subtribe Prasophyllinae: Several segregate genera are recognised in this subtribe [15, 23].
Within the recircumscribed Prasophyllum, most species have 2n = 42 (although P. rogersii from
Australia has 2n = 64). However, there is still a component of Prasophyllum with 2n = 44 and a
smaller chromosome complement (found in P. australe and P. brownii from Australia).

Corunastylis was segregated from Prasophyllum and there is support for this - New Zealand
material of Corunastylis nuda and C. pumila share a different chromosome number (2n = 44)
and have distinctly smaller chromosomes than most of Prasophyllum.

Mecopodum, an Australian segregate genus of Prasophyllum, has 2n = 44 (counted under its
earlier name, Prasophyllum parvifolium [18]). Microtis also has 2n = 44, except for the tetraploid
species M. unifolia. Again, further chromosome counts are needed and the other Australian genera
in subtribe Prasophyllinae remain uncounted.

Subtribe Thelymitrinae: Chromosome evidence strongly supports tribal treatments that accept
only Calochilus, Thelymitra, and the chromosomally unknown Epiblema (3, 4, 16; Fig. 1).

Chromosomes of Calochilus and Thelymitra are quite distinctive from other genera, and chromosome
evidence (where known) does not support Chase et al.'s [13] treatment that included 12 other genera.

Calochilus paludosus and C. robertsonii have 2n = 24, and C. aff. herbaceus of New Zealand has
2n = 22. The lower number is probably derived through aneuploidy within Calochilus. In turn,
2n = 24 may be derived from reduction of a 2n = 26 chromosome complement found in Thelymitra.

The wide range of chromosome numbers (2n = 26-93), limited aneuploidy, extensive allopolyploidy,
and reticulate evolution within Thelymitra was discussed in my previous article [1].

The occurrence of intergeneric hybrids between Calochilus and Thelymitra and similar chromosome
complements suggest that the two genera are closely related.

Subtribe Townsoniinae: This subtribe is made up of only one genus, Townsonia, and has been
separated from subtribe Acianthinae [4]. One of the two species is counted, T. deflexa with 2n = 28.

Both species (T. deflexa endemic to New Zealand and T. viridis endemic to Tasmania) were formerly
placed in Acianthus (under the one name, A. viridis), but the distinctive chromosome number and
long chromosomes characterising Townsonia provide independent support for the removal of the
species from Acianthus and from subtribe Acianthinae.

Concluding Remarks

All of these case studies show the usefulness of the new chromosome information. Admittedly,
many more counts are needed for the Australian species, and for some segregate genera there is
little (if any) difference between their chromosome morphology/number and the genus to which
they belonged. For these, their status can neither be confirmed nor rejected by the chromosome
evidence.

Examples where chromosome counts are relatively uninformative include Adelopetalum (2n = 36, 38) and Ichthyostomum (2n = 38), both segregates of Bulbophyllum, the most species-rich orchid genus.

The commonest chromosome number in Bulbophyllum is 2n = 38 (followed by 2n = 40).

Likewise, Winika (2n = 40) is a segregate of Dendrobium, and (like Bulbophyllum) is a species-rich
member of Subtribe Dendrobiinae with predominantly 2n = 38 and secondarily 2n = 40. In these
examples, macromorphology and DNA evidence is more helpful in supporting them as segregate
genera.

Chromosomes are relatively conservative characters, i.e., they do not generally change much over
time, so it is not uncommon to have closely related genera sharing the same count and chromosome
morphology. However, because they are conservative characters, major differences in chromosome
complements, such as those highlighted here, are quite significant from a taxonomic and a
phylogenetic viewpoint. The new chromosome information provides an independent set of
characters that should, in conjunction with traditional and molecular characters, assist in greater
long-term stability of the Australasian orchid names and classifications.

References

1. Dawson MI 2008. Chromosomes of New Zealand native orchids - part 1 of 2.
   New Zealand Native Orchid Group Journal 108: 16-20.
2. Dawson MI, Molloy BPJ, Beuzenberg EJ 2007. Contributions to a chromosome atlas of the
   New Zealand flora - 39. Orchidaceae. New Zealand Journal of Botany 45: 611-684.
3. Clements MA, Jones DL, Sharma IK, Nightingale ME, Garratt MJ, Fitzgerald KJ, Mackenzie AM,
   Molloy BPJ 2002. Phylogenetics of Diurideae (Orchidaceae) based on the Internal Transcribed
   Spacer (ITS) regions of nuclear ribosomal DNA. Lindleyana 17: 135-171.
4. Jones DL, Clements MA, Sharma IK, Mackenzie AM, Molloy BPJ 2002. Nomenclatural notes
   arising from studies into the tribe Diurideae (Orchidaceae). The Orchadian 13: 437-468.
5. Szlachetko DL 2001. Genera et Species Orchidalium 1. Polish Botanical Journal 46: 11-26.
6. Jones DL, Clements MA 2002. 1. A reassessment of Pterostylis R.Br. (Orchidaceae). In: Jones DL,
   Clements MA ed. A review of Pterostylis (Orchidaceae). Australian Orchid Research 4: 3-63.
7. Jones DL, Clements MA 2002. 2. A new classification of Pterostylis R.Br. (Orchidaceae). In: Jones DL,
   Clements MA ed. A review of Pterostylis (Orchidaceae). Australian Orchid Research 4: 64-124.
8. Jones DL, Clements MA, Molloy BPJ 2002. 3. A synopsis of the Subtribe Pterostylidinae. In: Jones DL, Clements MA ed. A review of Pterostylis (Orchidaceae). Australian Orchid Research 4: 125-144.
9. Jones DL, Clements MA, Sharma IK, Mackenzie AM 2001. A new classification of Caladenia R.Br. (Orchidaceae). The Orchadian 13: 388-420.
10. Hopper SD, Brown AP 2004. Robert Brown's Caladenia revisited, including a revision of its sister
    genera Cyanicula, Ericksonella and Pheladenia (Caladeniinae: Orchidaceae).
    Australian Systematic Botany 17: 171-240.
11. St George I 2007. The New Zealand orchids. New Zealand Native Orchid Group Journal 106: 18-27.
12. Dressler RL 1993. Phylogeny and classification of the orchid family. Cambridge University Press. 314 p.
13. Chase MW, Cameron KM, Barrett RL, Freudenstein JV 2003. DNA data and Orchidaceae systematics:
    a new phylogenetic classification. In: Dixon KW, Kell SP, Barrett RL, Cribb PJ ed. Orchid conservation.
    Kota Kinabalu, Sabah, Natural History Publications (Borneo). Pp. 69-89.
14. Pridgeon AM, Cribb PJ, Chase MW, Rasmussen FN ed. 2005. Genera Orchidacearum. Volume 4. Epidendroideae (Part 1). Oxford University Press. 672 p.
15. Jones DL, Clements MA, Sharma IK, Mackenzie AM, Molloy BPJ 2002. Nomenclatural notes arising
    from studies into the tribe Diurideae (Orchidaceae). The Orchadian 13: 437-468.
16. Clements MA, Jones DL 2002. Nomenclatural notes arising from studies into the Tribe Diurideae
    (Orchidaceae): additions and corrections. The Orchadian 13: 502-503.
17. Jones DL, Clements MA, Sharma IK, Mackenzie AM 2001. A new classification of Caladenia R.Br. (Orchidaceae). The Orchadian 13: 388-420.
18. Peakall R, James SH 1989. Chromosome numbers of some Australian terrestrial orchids. Lindleyana 4: 85-88.
19. Larsen K 1966. Studies in the Flora of Thailand. 40. Cytology of vascular plants. II.
    Dansk Botanisk Arkiv 23: 375-399.
20. Jones DL, Clements MA 2005. Miscellaneous nomenclatural notes and changes in Australian,
    New Guinea and New Zealand Orchidaceae. The Orchadian 15: 33-42.
21. Jones DL, Clements MA, Molloy BPJ 1997. Waireia, a new genus of Orchidaceae from New Zealand.
    The Orchadian 12: 282-287.
22. Ishida G, Sera T, Hashimoto K 1992. Chromosome numbers of some orchids from New Caledonia.
    Bulletin of the Hiroshima Botanical Garden 14: 47-50.
23. Jones DL, Clements MA 2004. Miscellaneous new species, new genera, reinstated genera and
    new combinations in Australian Orchidaceae. The Orchadian, Scientific Supplement 14: i-xvi.